Cineversity

Hi all,

I’m having a bit of a problem in R15 that I’ve been ignoring for a while - in fact I suspect I was having the issue in R14 as well.
When I render out multi pass to a tiff sequence - using the standard sRGB 2.1 profile and then suck my renders into AE they come in as 32bit with a linear profile embedded and they don’t look right unless I linearise the work space (just interpreting them to sRGB does not help).

Is this just something that has happened as colour management has evolved through cinema 4d that the only way to get an exact composite match from my multi pass is to use linear workflow? I like using it for some projects though it can be cumbersome for others and would be happier using just a straight sRGB colour managed non linearised for certain jobs without having to tweak all my imported layers to match the original render.

Does this make sense or am I missing something?

Thanks,

Ben

Hi Ben,

What bit depth do you use for the tiff? What Ae version? What is the Interpret Footage setting?

Do you can share a single tiff here, can be very small in resolution.

All the best

Sassi

Thanks Sassi,

I am rendering out 16bit tiff for both regular image and multi pass image.
I can see on closer inspection that the regular image render is coming through sRGB and looks fine in the correct colour profile - but all the multi pass channels seem to render in linear space and therefore require linear space or gamma 1.0 blending in ae to display them correctly.
I’m using AE CC. I’d like to use OpenEXR though its not as favoured by the facilities I work in.

Is there somewhere on cineversity I can upload a file or would you prefer a link?

Thanks,

Ben

Ok I have worked it out - silly me - I had linear workflow enabled by default in the project settings. Problem solved.

Thanks!

Hi Ben,

Nice that your found your solution. I wanted first check what data you have. A quick check in Ps is sometimes the fastest way. The color space sRGB is available in gamma as well in linear, so that was not clear to me where you got it from.

The linear workflow (Project Settings) overwrites the color profile from the Save As (single file). This might be not a tragic if the right profile is then set in Ae or Ps. However, the distribution in integer formats is limited, and so the mid gray is moved to 0.18, if linear is used (in float numbers), which squeezes the dark values and allows for more bright tones… nothing that I would generally suggest (perhaps for a high key image ;o).

Hence the suggestion in the manual to use at least 16bit/channel if float can’t be used while in linear mode. To use 8bit/c means here that the mid value is placed at 18% (if 100% is considered max). This is a fifth roughly of the available data for anything below mid gray. It gets even worse, as the distribution is doubled (or halved downwards) for each “stop” of brightness, anything below one stop under the mid value is then pretty much in a 1/10 of the 256 steps per channel. Two stops down everything else in roughly 5% of the available data and so on. That the “blacks” will suffer tremendously is clear, and any lifting of the values in a color-grading session will destroy the image. True as well for data and mask (alfa) channels.
As a side note, I highly recommend NOT to use at any time inside the workflow HDR-Radience (4*8bit. RGB+E) as a format, it is only (!) a 32bit/pixel and not per channel format.

All the best

Sassi